In this notebook, some template code has already been provided for you, and you will need to implement additional functionality to successfully complete this project. You will not need to modify the included code beyond what is requested. Sections that begin with '(IMPLEMENTATION)' in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section, and the specifics of the implementation are marked in the code block with a 'TODO' statement. Please be sure to read the instructions carefully!
Note: Once you have completed all of the code implementations, you need to finalize your work by exporting the iPython Notebook as an HTML document. Before exporting the notebook to html, all of the code cells need to have been run so that reviewers can see the final implementation and output. You can then export the notebook by using the menu above and navigating to \n", "File -> Download as -> HTML (.html). Include the finished document along with this notebook as your submission.
In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a 'Question X' header. Carefully read each question and provide thorough answers in the following text boxes that begin with 'Answer:'. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.
Note: Code and Markdown cells can be executed using the Shift + Enter keyboard shortcut. Markdown cells can be edited by double-clicking the cell to enter edit mode.
The rubric contains optional "Stand Out Suggestions" for enhancing the project beyond the minimum requirements. If you decide to pursue the "Stand Out Suggestions", you should include the code in this IPython notebook.
In this notebook, you will make the first steps towards developing an algorithm that could be used as part of a mobile or web app. At the end of this project, your code will accept any user-supplied image as input. If a dog is detected in the image, it will provide an estimate of the dog's breed. If a human is detected, it will provide an estimate of the dog breed that is most resembling. The image below displays potential sample output of your finished project (... but we expect that each student's algorithm will behave differently!).

In this real-world setting, you will need to piece together a series of models to perform different tasks; for instance, the algorithm that detects humans in an image will be different from the CNN that infers dog breed. There are many points of possible failure, and no perfect algorithm exists. Your imperfect solution will nonetheless create a fun user experience!
We break the notebook into separate steps. Feel free to use the links below to navigate the notebook.
In the code cell below, we import a dataset of dog images. We populate a few variables through the use of the load_files function from the scikit-learn library:
train_files, valid_files, test_files - numpy arrays containing file paths to imagestrain_targets, valid_targets, test_targets - numpy arrays containing onehot-encoded classification labels dog_names - list of string-valued dog breed names for translating labelsfrom sklearn.datasets import load_files
from keras.utils import np_utils
import numpy as np
from glob import glob
# define function to load train, test, and validation datasets
def load_dataset(path):
data = load_files(path)
dog_files = np.array(data['filenames'])
dog_targets = np_utils.to_categorical(np.array(data['target']), 133)
return dog_files, dog_targets
# load train, test, and validation datasets
train_files, train_targets = load_dataset('dogImages/train')
valid_files, valid_targets = load_dataset('dogImages/valid')
test_files, test_targets = load_dataset('dogImages/test')
# load list of dog names
dog_names = [item[20:-1] for item in sorted(glob("dogImages/train/*/"))]
# print statistics about the dataset
print('There are %d total dog categories.' % len(dog_names))
print('There are %s total dog images.\n' % len(np.hstack([train_files, valid_files, test_files])))
print('There are %d training dog images.' % len(train_files))
print('There are %d validation dog images.' % len(valid_files))
print('There are %d test dog images.'% len(test_files))
In the code cell below, we import a dataset of human images, where the file paths are stored in the numpy array human_files.
import random
random.seed(8675309)
# load filenames in shuffled human dataset
human_files = np.array(glob("lfw/*/*"))
random.shuffle(human_files)
# print statistics about the dataset
print('There are %d total human images.' % len(human_files))
We use OpenCV's implementation of Haar feature-based cascade classifiers to detect human faces in images. OpenCV provides many pre-trained face detectors, stored as XML files on github. We have downloaded one of these detectors and stored it in the haarcascades directory.
In the next code cell, we demonstrate how to use this detector to find human faces in a sample image.
import cv2
import matplotlib.pyplot as plt
%matplotlib inline
# extract pre-trained face detector
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')
# load color (BGR) image
img = cv2.imread(human_files[3])
# convert BGR image to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# find faces in image
faces = face_cascade.detectMultiScale(gray)
# print number of faces detected in the image
print('Number of faces detected:', len(faces))
# get bounding box for each detected face
for (x,y,w,h) in faces:
# add bounding box to color image
cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
# convert BGR image to RGB for plotting
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
# display the image, along with bounding box
plt.imshow(cv_rgb)
plt.show()
Before using any of the face detectors, it is standard procedure to convert the images to grayscale. The detectMultiScale function executes the classifier stored in face_cascade and takes the grayscale image as a parameter.
In the above code, faces is a numpy array of detected faces, where each row corresponds to a detected face. Each detected face is a 1D array with four entries that specifies the bounding box of the detected face. The first two entries in the array (extracted in the above code as x and y) specify the horizontal and vertical positions of the top left corner of the bounding box. The last two entries in the array (extracted here as w and h) specify the width and height of the box.
We can use this procedure to write a function that returns True if a human face is detected in an image and False otherwise. This function, aptly named face_detector, takes a string-valued file path to an image as input and appears in the code block below.
# returns "True" if face is detected in image stored at img_path
def face_detector(img_path):
img = cv2.imread(img_path)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray)
return len(faces) > 0
def visualize_detected_faces(img_path):
img = cv2.imread(img_path)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray)
# print number of faces detected in the image
print('Number of faces detected:', len(faces))
# get bounding box for each detected face
for (x,y,w,h) in faces:
# add bounding box to color image
cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
# convert BGR image to RGB for plotting
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
# display the image, along with bounding box
plt.imshow(cv_rgb)
plt.show()
Question 1: Use the code cell below to test the performance of the face_detector function.
human_files have a detected human face? dog_files have a detected human face? Ideally, we would like 100% of human images with a detected face and 0% of dog images with a detected face. You will see that our algorithm falls short of this goal, but still gives acceptable performance. We extract the file paths for the first 100 images from each of the datasets and store them in the numpy arrays human_files_short and dog_files_short.
Answer: In the first 100 images from the human_files, 99% include a detected human face. The one image that does not detect the face is turned to the side and does not provide enough features for the front face Haar cascades classifier used. In contrast, 11% of the first 100 images from the dog_files report a detected face. One of these 11 in fact includes a human in the photo, and the remaining 10 have areas that trigger false positive detections.
human_files_short = human_files[:100]
dog_files_short = train_files[:100]
# Do NOT modify the code above this line.
## TODO: Test the performance of the face_detector algorithm
## on the images in human_files_short and dog_files_short.
human_human_count = 0
dog_human_count = 0
humans_in_dog_images = []
no_humans_in_human_images = []
for i in range(len(human_files_short)):
if face_detector(human_files_short[i]):
human_human_count += 1
else:
no_humans_in_human_images.append(i)
if face_detector(dog_files_short[i]):
dog_human_count += 1
humans_in_dog_images.append(i)
print("{}% human faces detected in lfw".format(human_human_count))
print("{}% human faces detected in dogImages".format(dog_human_count))
for i in no_humans_in_human_images:
visualize_detected_faces(human_files_short[i])
for i in humans_in_dog_images:
visualize_detected_faces(dog_files_short[i])
Question 2: This algorithmic choice necessitates that we communicate to the user that we accept human images only when they provide a clear view of a face (otherwise, we risk having unneccessarily frustrated users!). In your opinion, is this a reasonable expectation to pose on the user? If not, can you think of a way to detect humans in images that does not necessitate an image with a clearly presented face?
Answer: Since the final application expects the user to upload a single photo at a time for analysis, instead of batch processing a large volume of images, the requirement to have a good view of the face should not be unreasonable. However, it was found that the OpenCV face detector is prone to false detections on photos of dogs alone. This could throw off any species specific greeting, so an alternative face detector will be evaluated.
Two alternatives were tested, both from the dlib Python library.
dlib.get_frontal_face_detector() which is based on the Histogram of Oriented Gradients with Linear SVM object detection method. This method performed on par with the OpenCV Haar cascades implementation, which detected faces in 99% of the first 100 human_files and 9% of the first 100 dog_files. The same side profile human image that failed in OpenCV failed the dlib detector which relies on a full frontal view of the face. The dlib detector found a second human face in the first 100 dog_files that OpenCV missed, and the other 7 images were false detections that were slightly different that the OpenCV false detections.human_files and 2% on the dog_files, where both images actually contain a human.We suggest the face detector from OpenCV as a potential way to detect human images in your algorithm, but you are free to explore other approaches, especially approaches that make use of deep learning :). Please use the code cell below to design and test your own face detection algorithm. If you decide to pursue this optional task, report performance on each of the datasets.
## (Optional) TODO: Report the performance of another
## face detection algorithm on the LFW dataset
### Feel free to use as many code cells as needed.
import dlib
detector = dlib.get_frontal_face_detector()
def face_detector(img_path):
img = cv2.imread(img_path)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = detector(gray, 1)
return len(faces) > 0
human_human_count = 0
dog_human_count = 0
humans_in_dog_images = []
no_humans_in_human_images = []
for i in range(len(human_files_short)):
if face_detector(human_files_short[i]):
human_human_count += 1
else:
no_humans_in_human_images.append(i)
if face_detector(dog_files_short[i]):
dog_human_count += 1
humans_in_dog_images.append(i)
print("{}% human faces detected in lfw".format(human_human_count))
print("{}% human faces detected in dogImages".format(dog_human_count))
for i in no_humans_in_human_images:
visualize_detected_faces(human_files_short[i])
for i in humans_in_dog_images:
visualize_detected_faces(dog_files_short[i])
detector = dlib.cnn_face_detection_model_v1('face_detection/mmod_human_face_detector.dat')
def face_detector(img_path):
img = cv2.imread(img_path)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = detector(gray, 1)
return len(faces) > 0
human_human_count = 0
dog_human_count = 0
humans_in_dog_images = []
no_humans_in_human_images = []
for i in range(len(human_files_short)):
if face_detector(human_files_short[i]):
human_human_count += 1
else:
no_humans_in_human_images.append(i)
if face_detector(dog_files_short[i]):
dog_human_count += 1
humans_in_dog_images.append(i)
print("{}% human faces detected in lfw".format(human_human_count))
print("{}% human faces detected in dogImages".format(dog_human_count))
In this section, we use a pre-trained ResNet-50 model to detect dogs in images. Our first line of code downloads the ResNet-50 model, along with weights that have been trained on ImageNet, a very large, very popular dataset used for image classification and other vision tasks. ImageNet contains over 10 million URLs, each linking to an image containing an object from one of 1000 categories. Given an image, this pre-trained ResNet-50 model returns a prediction (derived from the available categories in ImageNet) for the object that is contained in the image.
from keras.applications.resnet50 import ResNet50
# define ResNet50 model
ResNet50_model = ResNet50(weights='imagenet')
When using TensorFlow as backend, Keras CNNs require a 4D array (which we'll also refer to as a 4D tensor) as input, with shape
$$ (\text{nb_samples}, \text{rows}, \text{columns}, \text{channels}), $$where nb_samples corresponds to the total number of images (or samples), and rows, columns, and channels correspond to the number of rows, columns, and channels for each image, respectively.
The path_to_tensor function below takes a string-valued file path to a color image as input and returns a 4D tensor suitable for supplying to a Keras CNN. The function first loads the image and resizes it to a square image that is $224 \times 224$ pixels. Next, the image is converted to an array, which is then resized to a 4D tensor. In this case, since we are working with color images, each image has three channels. Likewise, since we are processing a single image (or sample), the returned tensor will always have shape
The paths_to_tensor function takes a numpy array of string-valued image paths as input and returns a 4D tensor with shape
Here, nb_samples is the number of samples, or number of images, in the supplied array of image paths. It is best to think of nb_samples as the number of 3D tensors (where each 3D tensor corresponds to a different image) in your dataset!
from keras.preprocessing import image
from tqdm import tqdm
def path_to_tensor(img_path):
# loads RGB image as PIL.Image.Image type
img = image.load_img(img_path, target_size=(224, 224))
# convert PIL.Image.Image type to 3D tensor with shape (224, 224, 3)
x = image.img_to_array(img)
# convert 3D tensor to 4D tensor with shape (1, 224, 224, 3) and return 4D tensor
return np.expand_dims(x, axis=0)
def paths_to_tensor(img_paths):
list_of_tensors = [path_to_tensor(img_path) for img_path in tqdm(img_paths)]
return np.vstack(list_of_tensors)
Getting the 4D tensor ready for ResNet-50, and for any other pre-trained model in Keras, requires some additional processing. First, the RGB image is converted to BGR by reordering the channels. All pre-trained models have the additional normalization step that the mean pixel (expressed in RGB as $[103.939, 116.779, 123.68]$ and calculated from all pixels in all images in ImageNet) must be subtracted from every pixel in each image. This is implemented in the imported function preprocess_input. If you're curious, you can check the code for preprocess_input here.
Now that we have a way to format our image for supplying to ResNet-50, we are now ready to use the model to extract the predictions. This is accomplished with the predict method, which returns an array whose $i$-th entry is the model's predicted probability that the image belongs to the $i$-th ImageNet category. This is implemented in the ResNet50_predict_labels function below.
By taking the argmax of the predicted probability vector, we obtain an integer corresponding to the model's predicted object class, which we can identify with an object category through the use of this dictionary.
from keras.applications.resnet50 import preprocess_input, decode_predictions
def ResNet50_predict_labels(img_path):
# returns prediction vector for image located at img_path
img = preprocess_input(path_to_tensor(img_path))
return np.argmax(ResNet50_model.predict(img))
While looking at the dictionary, you will notice that the categories corresponding to dogs appear in an uninterrupted sequence and correspond to dictionary keys 151-268, inclusive, to include all categories from 'Chihuahua' to 'Mexican hairless'. Thus, in order to check to see if an image is predicted to contain a dog by the pre-trained ResNet-50 model, we need only check if the ResNet50_predict_labels function above returns a value between 151 and 268 (inclusive).
We use these ideas to complete the dog_detector function below, which returns True if a dog is detected in an image (and False if not).
### returns "True" if a dog is detected in the image stored at img_path
def dog_detector(img_path):
prediction = ResNet50_predict_labels(img_path)
return ((prediction <= 268) & (prediction >= 151))
Question 3: Use the code cell below to test the performance of your dog_detector function.
human_files_short have a detected dog? dog_files_short have a detected dog?Answer: The dog_detector function generates 1 false detection in the first 100 human_files that does not have a dog in it, and detects a dog in 100% of the first 100 dog_files. This achieves the goal of a dog detection algorithm.
### TODO: Test the performance of the dog_detector function
### on the images in human_files_short and dog_files_short.
dog_human_count = 0
dog_dog_count = 0
dog_in_human_images = []
no_dog_in_dog_images = []
for i in range(len(human_files_short)):
if dog_detector(human_files_short[i]):
dog_human_count += 1
dog_in_human_images.append(i)
if dog_detector(dog_files_short[i]):
dog_dog_count += 1
else:
no_dog_in_dog_images.append(i)
print("{}% dogs detected in sample lfw".format(dog_human_count))
print("{}% dogs detected in sample dogImages".format(dog_dog_count))
visualize_detected_faces(human_files_short[dog_in_human_images[0]])
Now that we have functions for detecting humans and dogs in images, we need a way to predict breed from images. In this step, you will create a CNN that classifies dog breeds. You must create your CNN from scratch (so, you can't use transfer learning yet!), and you must attain a test accuracy of at least 1%. In Step 5 of this notebook, you will have the opportunity to use transfer learning to create a CNN that attains greatly improved accuracy.
Be careful with adding too many trainable layers! More parameters means longer training, which means you are more likely to need a GPU to accelerate the training process. Thankfully, Keras provides a handy estimate of the time that each epoch is likely to take; you can extrapolate this estimate to figure out how long it will take for your algorithm to train.
We mention that the task of assigning breed to dogs from images is considered exceptionally challenging. To see why, consider that even a human would have great difficulty in distinguishing between a Brittany and a Welsh Springer Spaniel.
| Brittany | Welsh Springer Spaniel |
|---|---|
![]() |
![]() |
It is not difficult to find other dog breed pairs with minimal inter-class variation (for instance, Curly-Coated Retrievers and American Water Spaniels).
| Curly-Coated Retriever | American Water Spaniel |
|---|---|
![]() |
![]() |
Likewise, recall that labradors come in yellow, chocolate, and black. Your vision-based algorithm will have to conquer this high intra-class variation to determine how to classify all of these different shades as the same breed.
| Yellow Labrador | Chocolate Labrador | Black Labrador |
|---|---|---|
![]() |
![]() |
![]() |
We also mention that random chance presents an exceptionally low bar: setting aside the fact that the classes are slightly imabalanced, a random guess will provide a correct answer roughly 1 in 133 times, which corresponds to an accuracy of less than 1%.
Remember that the practice is far ahead of the theory in deep learning. Experiment with many different architectures, and trust your intuition. And, of course, have fun!
We rescale the images by dividing every pixel in every image by 255.
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True
# pre-process the data for Keras
train_tensors = paths_to_tensor(train_files).astype('float32')/255
valid_tensors = paths_to_tensor(valid_files).astype('float32')/255
test_tensors = paths_to_tensor(test_files).astype('float32')/255
Create a CNN to classify dog breed. At the end of your code cell block, summarize the layers of your model by executing the line:
model.summary()
We have imported some Python modules to get you started, but feel free to import as many modules as you need. If you end up getting stuck, here's a hint that specifies a model that trains relatively fast on CPU and attains >1% test accuracy in 5 epochs:

Question 4: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. If you chose to use the hinted architecture above, describe why you think that CNN architecture should work well for the image classification task.
Answer:
First attempted network achieved about 6.5% accuracy on the test set without data augmentation, and about 8% with augmentation after 10 epochs with batch size of 20.
The following architecture was used:
model.add(Conv2D(128, kernel_size=3, padding='same', input_shape=(224, 224,3)))
model.add(Activation('relu'))
model.add(Dropout(0.2))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
model.add(Conv2D(64, kernel_size=3, padding='same'))
model.add(Activation('relu'))
model.add(Dropout(0.2))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
model.add(Conv2D(32, kernel_size=3, padding='same'))
model.add(Activation('relu'))
model.add(Dropout(0.1))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
model.add(Flatten())
model.add(Dense(len(dog_names)))
model.add(Activation('softmax'))
model.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d_8 (Conv2D) (None, 224, 224, 128) 3584
_________________________________________________________________
activation_58 (Activation) (None, 224, 224, 128) 0
_________________________________________________________________
dropout_7 (Dropout) (None, 224, 224, 128) 0
_________________________________________________________________
max_pooling2d_8 (MaxPooling2 (None, 112, 112, 128) 0
_________________________________________________________________
conv2d_9 (Conv2D) (None, 112, 112, 64) 73792
_________________________________________________________________
activation_59 (Activation) (None, 112, 112, 64) 0
_________________________________________________________________
dropout_8 (Dropout) (None, 112, 112, 64) 0
_________________________________________________________________
max_pooling2d_9 (MaxPooling2 (None, 56, 56, 64) 0
_________________________________________________________________
conv2d_10 (Conv2D) (None, 56, 56, 32) 18464
_________________________________________________________________
activation_60 (Activation) (None, 56, 56, 32) 0
_________________________________________________________________
dropout_9 (Dropout) (None, 56, 56, 32) 0
_________________________________________________________________
max_pooling2d_10 (MaxPooling (None, 28, 28, 32) 0
_________________________________________________________________
flatten_4 (Flatten) (None, 25088) 0
_________________________________________________________________
dense_3 (Dense) (None, 133) 3336837
_________________________________________________________________
activation_61 (Activation) (None, 133) 0
=================================================================
Total params: 3,432,677.0
Trainable params: 3,432,677.0
Non-trainable params: 0.0
_________________________________________________________________
The second model attempted replaced the Flatten() layer with a GlobalAveragePooling2D() layer just before the final dense layer. This reduced the total number of parameters from 3,432,677 to 404,997 which cut down on training time. The number of filters in each Conv2D layer was also reversed to double each time instead of decreasing by half.
With data augmentation and 10 epochs of batch size 32, the model achieved Test accuracy: 7.0574%. After training the model for another 10 epochs, the model achieved Test accuracy: 13.0383%.
Architecture is as follows:
model.add(Conv2D(64, kernel_size=3, padding='same', input_shape=(224, 224,3)))
model.add(Activation('relu'))
model.add(Dropout(0.3))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
model.add(Conv2D(128, kernel_size=3, padding='same'))
model.add(Activation('relu'))
model.add(Dropout(0.2))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
model.add(Conv2D(256, kernel_size=3, padding='same'))
model.add(Activation('relu'))
model.add(Dropout(0.1))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
#model.add(Flatten())
model.add(GlobalAveragePooling2D())
model.add(Dense(len(dog_names)))
model.add(Activation('softmax'))
model.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d_11 (Conv2D) (None, 224, 224, 64) 1792
_________________________________________________________________
activation_62 (Activation) (None, 224, 224, 64) 0
_________________________________________________________________
dropout_10 (Dropout) (None, 224, 224, 64) 0
_________________________________________________________________
max_pooling2d_11 (MaxPooling (None, 112, 112, 64) 0
_________________________________________________________________
conv2d_12 (Conv2D) (None, 112, 112, 128) 73856
_________________________________________________________________
activation_63 (Activation) (None, 112, 112, 128) 0
_________________________________________________________________
dropout_11 (Dropout) (None, 112, 112, 128) 0
_________________________________________________________________
max_pooling2d_12 (MaxPooling (None, 56, 56, 128) 0
_________________________________________________________________
conv2d_13 (Conv2D) (None, 56, 56, 256) 295168
_________________________________________________________________
activation_64 (Activation) (None, 56, 56, 256) 0
_________________________________________________________________
dropout_12 (Dropout) (None, 56, 56, 256) 0
_________________________________________________________________
max_pooling2d_13 (MaxPooling (None, 28, 28, 256) 0
_________________________________________________________________
global_average_pooling2d_1 ( (None, 256) 0
_________________________________________________________________
dense_4 (Dense) (None, 133) 34181
_________________________________________________________________
activation_65 (Activation) (None, 133) 0
=================================================================
Total params: 404,997.0
Trainable params: 404,997.0
Non-trainable params: 0.0
_________________________________________________________________
from keras.layers import Conv2D, MaxPooling2D, GlobalAveragePooling2D
from keras.layers import Dropout, Flatten, Dense, Activation
from keras.models import Sequential
model = Sequential()
### TODO: Define your architecture.
model.add(Conv2D(64, kernel_size=3, padding='same', input_shape=(224, 224,3)))
model.add(Activation('relu'))
model.add(Dropout(0.3))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
model.add(Conv2D(128, kernel_size=3, padding='same'))
model.add(Activation('relu'))
model.add(Dropout(0.2))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
model.add(Conv2D(256, kernel_size=3, padding='same'))
model.add(Activation('relu'))
model.add(Dropout(0.1))
model.add(MaxPooling2D(pool_size=(2,2), padding='same'))
#model.add(Flatten())
model.add(GlobalAveragePooling2D())
model.add(Dense(len(dog_names)))
model.add(Activation('softmax'))
model.summary()
model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])
Train your model in the code cell below. Use model checkpointing to save the model that attains the best validation loss.
You are welcome to augment the training data, but this is not a requirement.
from keras.preprocessing.image import ImageDataGenerator
datagen_train = ImageDataGenerator(
width_shift_range=0.2,
height_shift_range=0.2,
rotation_range=5,
horizontal_flip=True)
datagen_valid = ImageDataGenerator(
width_shift_range=0.2,
height_shift_range=0.2,
rotation_range=5,
horizontal_flip=True)
datagen_train.fit(train_tensors)
datagen_valid.fit(valid_tensors)
from keras.callbacks import ModelCheckpoint
### TODO: specify the number of epochs that you would like to use to train the model.
epochs = 20
batch_size = 32
### Do NOT modify the code below this line.
checkpointer = ModelCheckpoint(filepath='saved_models/weights.best.from_scratch.hdf5',
verbose=1, save_best_only=True)
#model.fit(train_tensors, train_targets,
# validation_data=(valid_tensors, valid_targets),
# epochs=epochs, batch_size=20, callbacks=[checkpointer], verbose=1)
model.fit_generator(datagen_train.flow(train_tensors, train_targets, batch_size=batch_size),
steps_per_epoch=train_tensors.shape[0] // batch_size,
validation_data=datagen_valid.flow(
valid_tensors, valid_targets, batch_size=batch_size),
validation_steps=valid_tensors.shape[0] // batch_size,
epochs=epochs, callbacks=[checkpointer], verbose=2)
model.save_weights('saved_models/weights.last.from_scratch.hdf5')
model.load_weights('saved_models/weights.best.from_scratch.hdf5')
Try out your model on the test dataset of dog images. Ensure that your test accuracy is greater than 1%.
# get index of predicted dog breed for each image in test set
dog_breed_predictions = [np.argmax(model.predict(np.expand_dims(tensor, axis=0))) for tensor in test_tensors]
# report test accuracy
test_accuracy = 100*np.sum(np.array(dog_breed_predictions)==np.argmax(test_targets, axis=1))/len(dog_breed_predictions)
print('Test accuracy: %.4f%%' % test_accuracy)
bottleneck_features = np.load('bottleneck_features/DogVGG16Data.npz')
train_VGG16 = bottleneck_features['train']
valid_VGG16 = bottleneck_features['valid']
test_VGG16 = bottleneck_features['test']
The model uses the the pre-trained VGG-16 model as a fixed feature extractor, where the last convolutional output of VGG-16 is fed as input to our model. We only add a global average pooling layer and a fully connected layer, where the latter contains one node for each dog category and is equipped with a softmax.
VGG16_model = Sequential()
VGG16_model.add(GlobalAveragePooling2D(input_shape=train_VGG16.shape[1:]))
VGG16_model.add(Dense(133, activation='softmax'))
VGG16_model.summary()
VGG16_model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])
checkpointer = ModelCheckpoint(filepath='saved_models/weights.best.VGG16.hdf5',
verbose=1, save_best_only=True)
VGG16_model.fit(train_VGG16, train_targets,
validation_data=(valid_VGG16, valid_targets),
epochs=20, batch_size=20, callbacks=[checkpointer], verbose=1)
VGG16_model.save_weights('saved_models/weights.last.VGG16.hdf5')
VGG16_model.load_weights('saved_models/weights.best.VGG16.hdf5')
Now, we can use the CNN to test how well it identifies breed within our test dataset of dog images. We print the test accuracy below.
# get index of predicted dog breed for each image in test set
VGG16_predictions = [np.argmax(VGG16_model.predict(np.expand_dims(feature, axis=0))) for feature in test_VGG16]
# report test accuracy
test_accuracy = 100*np.sum(np.array(VGG16_predictions)==np.argmax(test_targets, axis=1))/len(VGG16_predictions)
print('Test accuracy: %.4f%%' % test_accuracy)
from extract_bottleneck_features import *
def VGG16_predict_breed(img_path):
# extract bottleneck features
bottleneck_feature = extract_VGG16(path_to_tensor(img_path))
# obtain predicted vector
predicted_vector = VGG16_model.predict(bottleneck_feature)
# return dog breed that is predicted by the model
return dog_names[np.argmax(predicted_vector)]
You will now use transfer learning to create a CNN that can identify dog breed from images. Your CNN must attain at least 60% accuracy on the test set.
In Step 4, we used transfer learning to create a CNN using VGG-16 bottleneck features. In this section, you must use the bottleneck features from a different pre-trained model. To make things easier for you, we have pre-computed the features for all of the networks that are currently available in Keras:
The files are encoded as such:
Dog{network}Data.npz
where {network}, in the above filename, can be one of VGG19, Resnet50, InceptionV3, or Xception. Pick one of the above architectures, download the corresponding bottleneck features, and store the downloaded file in the bottleneck_features/ folder in the repository.
In the code block below, extract the bottleneck features corresponding to the train, test, and validation sets by running the following:
bottleneck_features = np.load('bottleneck_features/Dog{network}Data.npz')
train_{network} = bottleneck_features['train']
valid_{network} = bottleneck_features['valid']
test_{network} = bottleneck_features['test']
### TODO: Obtain bottleneck features from another pre-trained CNN.
#bottleneck_features = np.load('bottleneck_features/DogVGG19Data.npz')
#train_VGG19 = bottleneck_features['train']
#valid_VGG19 = bottleneck_features['valid']
#test_VGG19 = bottleneck_features['test']
### TODO: Obtain bottleneck features from another pre-trained CNN.
#bottleneck_features = np.load('bottleneck_features/DogResnet50Data.npz')
#train_Resnet50 = bottleneck_features['train']
#valid_Resnet50 = bottleneck_features['valid']
#test_Resnet50 = bottleneck_features['test']
### TODO: Obtain bottleneck features from another pre-trained CNN.
#bottleneck_features = np.load('bottleneck_features/DogInceptionV3Data.npz')
#train_InceptionV3 = bottleneck_features['train']
#valid_InceptionV3 = bottleneck_features['valid']
#test_InceptionV3 = bottleneck_features['test']
### TODO: Obtain bottleneck features from another pre-trained CNN.
bottleneck_features = np.load('bottleneck_features/DogXceptionData.npz')
train_Xception = bottleneck_features['train']
valid_Xception = bottleneck_features['valid']
test_Xception = bottleneck_features['test']
train_Xception.shape
Create a CNN to classify dog breed. At the end of your code cell block, summarize the layers of your model by executing the line:
<your model's name>.summary()
Question 5: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. Describe why you think the architecture is suitable for the current problem.
Answer: Since the bottleneck features in the Xception dataset were trained against Imagenet which is generally includes quite similar images to the dog and human dataset used here, fine tuning should only require a final fully connected layer trained against the dogImages. In order to reduce the number of parameters, a GlobalAveragePooling2D layer was used. The test accuracy with 20 epochs is about 84%.
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
global_average_pooling2d_9 ( (None, 2048) 0
_________________________________________________________________
dense_10 (Dense) (None, 133) 272517
=================================================================
Total params: 272,517.0
Trainable params: 272,517.0
Non-trainable params: 0.0
_________________________________________________________________
The following architectures were also tested:
Adding an extra convolutional layer before the fully connected layer resulted in slightly lower test accuracy around 80% and increased the number of parameters several times over.
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d_12 (Conv2D) (None, 7, 7, 256) 4718848
_________________________________________________________________
activation_62 (Activation) (None, 7, 7, 256) 0
_________________________________________________________________
dropout_11 (Dropout) (None, 7, 7, 256) 0
_________________________________________________________________
max_pooling2d_12 (MaxPooling (None, 3, 3, 256) 0
_________________________________________________________________
global_average_pooling2d_8 ( (None, 256) 0
_________________________________________________________________
dense_9 (Dense) (None, 133) 34181
=================================================================
Total params: 4,753,029.0
Trainable params: 4,753,029.0
Non-trainable params: 0.0
_________________________________________________________________
Using a Flatten layer instead of the GlobalAveragePooling2D not only resulted in millions more parameters, but the test accuracy after 10 epochs was about 33%.
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
flatten_3 (Flatten) (None, 100352) 0
_________________________________________________________________
dense_11 (Dense) (None, 133) 13346949
=================================================================
Total params: 13,346,949.0
Trainable params: 13,346,949.0
Non-trainable params: 0.0
_________________________________________________________________
### TODO: Define your architecture.
Xception_model = Sequential()
Xception_model.add(GlobalAveragePooling2D(input_shape=train_Xception.shape[1:]))
Xception_model.add(Dense(133, activation='softmax'))
Xception_model.summary()
'''
Xception_model = Sequential()
Xception_model.add(Conv2D(256, kernel_size=3, padding='same', input_shape=train_Xception.shape[1:]))
Xception_model.add(Activation('relu'))
Xception_model.add(Dropout(0.4))
Xception_model.add(MaxPooling2D(pool_size=(3,3), padding='same'))
Xception_model.add(GlobalAveragePooling2D())
#Xception_model.add(Flatten())
Xception_model.add(Dense(133, activation='softmax'))
Xception_model.summary()
'''
'''
Xception_model = Sequential()
Xception_model.add(Flatten(input_shape=train_Xception.shape[1:]))
Xception_model.add(Dense(133, activation='softmax'))
Xception_model.summary()
'''
Xception_model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])
Train your model in the code cell below. Use model checkpointing to save the model that attains the best validation loss.
You are welcome to augment the training data, but this is not a requirement.
### TODO: Train the model.
epochs = 20
batch_size = 64
checkpointer = ModelCheckpoint(filepath='saved_models/weights.best.Xception.hdf5',
verbose=1, save_best_only=True)
Xception_model.fit(train_Xception, train_targets,
validation_data=(valid_Xception, valid_targets),
epochs=epochs, batch_size=batch_size, callbacks=[checkpointer], verbose=1)
Xception_model.save_weights('saved_models/weights.last.Xception.hdf5')
### TODO: Load the model weights with the best validation loss.
Xception_model.load_weights('saved_models/weights.best.Xception.hdf5')
Try out your model on the test dataset of dog images. Ensure that your test accuracy is greater than 60%.
### TODO: Calculate classification accuracy on the test dataset.
# get index of predicted dog breed for each image in test set
Xception_predictions = [np.argmax(Xception_model.predict(np.expand_dims(feature, axis=0))) for feature in test_Xception]
# report test accuracy
test_accuracy = 100*np.sum(np.array(Xception_predictions)==np.argmax(test_targets, axis=1))/len(Xception_predictions)
print('Test accuracy: %.4f%%' % test_accuracy)
InceptionV3: Test accuracy: 78.9474%
Xception: Test accuracy: 84.4498%
Resnet50: Test accuracy: 81.5789%
VGG19: Test accuracy: 36.4833%
Write a function that takes an image path as input and returns the dog breed (Affenpinscher, Afghan_hound, etc) that is predicted by your model.
Similar to the analogous function in Step 5, your function should have three steps:
dog_names array defined in Step 0 of this notebook to return the corresponding breed.The functions to extract the bottleneck features can be found in extract_bottleneck_features.py, and they have been imported in an earlier code cell. To obtain the bottleneck features corresponding to your chosen CNN architecture, you need to use the function
extract_{network}
where {network}, in the above filename, should be one of VGG19, Resnet50, InceptionV3, or Xception.
### TODO: Write a function that takes a path to an image as input
### and returns the dog breed that is predicted by the model.
def Xception_predict_breed(img_path):
# extract bottleneck features
bottleneck_feature = extract_Xception(path_to_tensor(img_path))
# obtain predicted vector
predicted_vector = Xception_model.predict(bottleneck_feature)
# return dog breed that is predicted by the model
return dog_names[np.argmax(predicted_vector)]
Write an algorithm that accepts a file path to an image and first determines whether the image contains a human, dog, or neither. Then,
You are welcome to write your own functions for detecting humans and dogs in images, but feel free to use the face_detector and dog_detector functions developed above. You are required to use your CNN from Step 5 to predict dog breed.
Some sample output for our algorithm is provided below, but feel free to design your own user experience!

### TODO: Write your algorithm.
### Feel free to use as many code cells as needed.
from extract_bottleneck_features import extract_Xception
from keras.layers import GlobalAveragePooling2D
from keras.layers import Dense
from keras.models import Sequential
import numpy as np
from glob import glob
# load list of dog names
dog_names = [item[20:-1] for item in sorted(glob("dogImages/train/*/"))]
Xception_model = Sequential()
Xception_model.add(GlobalAveragePooling2D(input_shape=(7, 7, 2048)))
Xception_model.add(Dense(133, activation='softmax'))
# Xception_model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])
Xception_model.load_weights('saved_models/weights.best.Xception.hdf5')
def Xception_predict_breed(img_path):
# extract bottleneck features
bottleneck_feature = extract_Xception(path_to_tensor(img_path))
# obtain predicted vector
predicted_vector = Xception_model.predict(bottleneck_feature)
# check the confidence of the predicted breed
top_3_predictions = decode_top_predictions(predicted_vector, top=3)
# return dog breed that is predicted by the model
return dog_names[np.argmax(predicted_vector)], top_3_predictions
def decode_top_predictions(preds, top=3):
result = []
for pred in preds:
top_indices = pred.argsort()[-top:][::-1]
for i in top_indices:
result.append([dog_names[i], pred[i]])
return result
from keras.applications.resnet50 import ResNet50, preprocess_input
from keras.preprocessing import image
def path_to_tensor(img_path):
# loads RGB image as PIL.Image.Image type
img = image.load_img(img_path, target_size=(224, 224))
# convert PIL.Image.Image type to 3D tensor with shape (224, 224, 3)
x = image.img_to_array(img)
# convert 3D tensor to 4D tensor with shape (1, 224, 224, 3) and return 4D tensor
return np.expand_dims(x, axis=0)
# define ResNet50 model
ResNet50_model = ResNet50(weights='imagenet')
def ResNet50_predict_labels(img_path):
# returns prediction vector for image located at img_path
img = preprocess_input(path_to_tensor(img_path))
preds = ResNet50_model.predict(img)
return np.argmax(preds)
def dog_detector(img_path):
prediction = ResNet50_predict_labels(img_path)
return ((prediction <= 268) & (prediction >= 151))
import cv2
import dlib
import matplotlib.pyplot as plt
%matplotlib inline
def display_image(img_path):
img = cv2.imread(img_path)
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
plt.imshow(cv_rgb)
plt.show()
# detector = dlib.cnn_face_detection_model_v1('face_detection/mmod_human_face_detector.dat')
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')
def face_detector(img_path):
img = cv2.imread(img_path)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray)
# faces = detector(gray, 1)
return len(faces) > 0
def dog_breed_detector(img_path):
has_dog = dog_detector(img_path)
has_face = face_detector(img_path)
if has_dog or has_face:
if has_dog:
print("Hey Doggy!")
display_image(img_path)
breed, top_3 = Xception_predict_breed(img_path)
if top_3[0][1] > 0.8:
print("I'm {}% sure this looks like a {}.".format(round(top_3[0][1]*100.,2),breed))
elif top_3[0][1] > 0.6:
print("I'm {}% sure this looks like a {}, but it could also be a {}.".format(
round(top_3[0][1]*100.,2),breed,top_3[1][0]))
else:
print("Hmmm...I'm not quite sure about this one! Here's my top 3 guesses:")
print("Prob.\tBreed")
for i in range(3):
print("{}\t{}".format(round(top_3[i][1]*100.,2), top_3[i][0]))
elif has_face:
print("Hello Human!")
display_image(img_path)
breed, top_3 = Xception_predict_breed(img_path)
print("You look {}% like a {}.".format(round(top_3[0][1]*100.,2),breed))
else:
print("Sorry! I didn't find any dogs or humans in your photo. Try again?")
display_image(img_path)
In this section, you will take your new algorithm for a spin! What kind of dog does the algorithm think that you look like? If you have a dog, does it predict your dog's breed accurately? If you have a cat, does it mistakenly think that your cat is a dog?
Test your algorithm at least six images on your computer. Feel free to use any images you like. Use at least two human and two dog images.
Question 6: Is the output better than you expected :) ? Or worse :( ? Provide at least three possible points of improvement for your algorithm.
Answer: The algorithm does surprisingly well. I don't have a dog of my own, but the photos of my friends' dogs are almost all identified correctly, and most with very high levels of confidence. The main exception seemed to be the long haired chihuahua which was classified as an Icelandic_sheepdog. Looking through the training data set, most of the Chihuahua images are short haired, so this isn't entirely surprising. While the dlib CNN based face detector performed much better than the OpenCV Haar cascades implementation, it generated CUDA out of memory errors on some photos, so the OpenCV face detector was used instead. The face and dog detectors work well enough that a photo of a cat doesn't undergo analysis.
## TODO: Execute your algorithm from Step 6 on
## at least 6 images on your computer.
## Feel free to use as many code cells as needed.
dog_breed_detector('images/861A8795.jpg')
dog_breed_detector('images/3250667833_df75e92aec_z_d.jpg')
Check what breed of dog I look like.
dog_breed_detector('images/Photo_on_2018-01-19_at_3_20_PM.jpg')
dog_breed_detector('images/637e8ed400a69b54c17c1402a1912d7f.jpg')
dog_breed_detector('images/ricky-gervais.jpg')
dog_breed_detector('images/40.jpg')
dog_breed_detector('images/IMG_1041.JPG')
dog_breed_detector('images/IMG_2491.JPG')
dog_breed_detector('images/IMG_4255.jpg')
dog_breed_detector('images/IMG_4415.PNG')
dog_breed_detector('images/Labrador_retriever_06449.jpg')